: Heart failure (HF) has reached epidemic proportions effecting over 5.4 million people worldwide and its burden is expected to grow exponentially in the next decade. Mechanical circulatory support (MCS) with continuous flow (CF) pumps for patients with end stage heart failure or acute cardiogenic shock has led to drastic improvements in survival. However, device operation remains fraught with serious adverse events (SAEs) leading to excessive morbidity. CF pump thrombosis and stroke are occurring at high rates, thereby prompting a need to uncover their hematologic mechanisms and therapeutic targets. In the chronic CF pump field, the newly designed Heart Mate (HM) 3 device, with features of greater hemocompatability such as wider blood flow passages, did show a reduction of in-situ pump thrombosis; however, the presence of disabling ischemic stroke remained elevated. Such persistence of severe thrombotic events, despite alterations in device design, makes it critically important to fully elucidate mechanisms underlying thrombus formation. The association of hemolysis with device thrombosis and ischemic stroke has recently been recognized and is consistent with a hemolysis induced pro-thrombotic state. Since overt clinical hemolysis may be a downstream effect of thrombus already formed in the device and given that in-situ thrombosis may no longer occur in newer generation of durable devices, we instead focus on assessing the in-vivo impact of sub clinical low level hemolysis (LLH) on platelet activation and aggregation (PA/A). In hemolytic disorders such as sickle cell anemia, extracellular hemoglobin may induce PA/A by scavenging nitric oxide (NO) and binding to pro- thrombotic ultra large (UL) multimers of von Willebrand factor (vWF) to prevent their proteolysis by the metalloproteinase (ADAMTS 13). Our preliminary findings indicate that during both chronic and acute CF MCS, LLH was associated with a 7 to 15 fold increased risk of subsequent thrombotic events, including ischemic stroke. Moreover, we noted that mean platelet volume (MPV), a surrogate marker of PA/A, and pro-thrombotic vWF levels were higher in patients with LLH in comparison to those with no LLH. Our findings also indicate a substantial reduction in thrombotic events and MPV with ongoing LLH during durable CF pump support in patients receiving a platelet NO signaling potentiation agent, sildenafil. This proposal will evaluate the hypotheses that LLH during CF pump support is associated with an increase in PA/A and sildenafil can lower PA/A during LLH. In Aim 1, we will recruit patients on chronic (HM 2 and HM 3) and acute (VA ECMO and Impella) CF MCS, to assess PA/A between those with and without LLH. In secondary analyses, we will also examine whether this association may be related to changes in platelet cGMP, vWF activity, vWF UL/HMW multimers and ADAMTS13 activity. In Aim 2, outpatients on durable CF pump support with ongoing LLH will be randomized (1:1) to receive sildenafil or placebo in a double-blinded manner for a 15 day period and changes in PA/A markers will be measured after assigned drug administration and compared to baseline.